Neurocognitive therapeutics: from concept to application in the treatment of negative attention bias

David M. Schnyer, Christopher G. Beevers, Megan T. deBettencourt, Stephanie M. Sherman, Jonathan D. Cohen, Kenneth A. Norman, Nicholas B. Turk‐BrowneBiology 2015 · DOI

TL;DR

This paper introduces "neurocognitive therapeutics," a novel framework that uses real-time fMRI and machine learning to dynamically adjust cognitive tasks based on an individual's brain activity. A pilot study demonstrated its potential to alter negative attention bias in depressed individuals, leading to improved mood symptoms and changes in brain connectivity.

What Problem Does This Paper Address?

Mental illnesses, particularly mood disorders, are often linked to maladaptive cognitive patterns, such as an excessive focus on negative stimuli (negative attention bias). Traditional cognitive training aims to retrain these patterns, but its effectiveness can be limited by a one-size-fits-all approach. The authors propose a more personalized method, "neurocognitive therapeutics," which leverages the precision of neuroimaging and machine learning to tailor interventions in real-time, adapting the cognitive task to the individual's current neural state. This aims to overcome the limitations of behavioral feedback alone and enhance neuroplasticity for more effective treatment.

Methods

The core innovation is the combination of real-time functional Magnetic Resonance Imaging (fMRI) with Multivoxel Pattern Analysis (MVPA). Here's how it works:

  1. Classifier Training: An MVPA classifier is trained offline using fMRI data from participants performing a task. In this case, participants were asked to attend to either faces or scenes within composite images, and the classifier learned to distinguish brain activity patterns associated with attending to each category.
  2. Real-time Decoding & Feedback: During the actual training session, fMRI data is acquired and processed in real-time. The trained classifier then decodes the participant's current attentional state (e.g., whether they are attending to scenes or being distracted by sad faces).
  3. Dynamic Task Adaptation: Based on the real-time decoding, the visual display is dynamically altered. If the classifier detects that the participant is being distracted by negative stimuli (sad faces), the task is made more challenging in a way that encourages re-engagement with the target stimulus (scenes). For instance, the proportion of the scene in the composite image might be reduced. This provides a tangible, immediate "error signal" to guide the participant's attention.

This "closed-loop" system allows for continuous adaptation of the cognitive task, aiming to optimize the training for each individual's specific needs and neural responses.

Datasets

The study reports on a pilot feasibility trial involving seven adults with elevated symptoms of depression. The data used for training the MVPA classifier and for the real-time feedback sessions were collected from these participants. No external datasets were explicitly mentioned for training or evaluation. The study's design is novel in its application of real-time fMRI neurofeedback to target negative attention bias, rather than relying on pre-existing benchmarks for this specific therapeutic approach.

Key Insights and Ideas

  • Personalized Cognitive Training: The most significant insight is the potential for neuroimaging and machine learning to create truly personalized cognitive interventions. By decoding brain states in real-time, treatments can be dynamically adjusted, moving beyond generic behavioral feedback.
  • "Externalizing" Mental States: Providing real-time feedback based on brain activity can make abstract cognitive processes, like attentional bias, more tangible and understandable to the individual. This "externalization" can facilitate learning and self-regulation.
  • Bridging Cognitive Neuroscience and Therapeutics: The paper highlights a promising pathway for translating cutting-edge neuroscience techniques into clinical applications, offering a new paradigm for treating mental health conditions.
  • Attention as a Therapeutic Target: The study reinforces the idea that improving attentional control is a viable mechanism for alleviating mood symptoms, and that this control can be enhanced through targeted interventions.

Key Results

The pilot feasibility study yielded promising preliminary results:

  • Feasibility: The complex real-time fMRI neurofeedback procedure was successfully implemented in patients with depression.
  • Mood Improvement: Improvements in attention control during training predicted subsequent improvements in depression symptoms over a 4-week follow-up period.
  • Brain Connectivity Changes: Resting-state fMRI revealed increased connectivity within a previously identified attention control network (specifically between the right middle frontal gyrus and bilateral supramarginal gyrus) from before to after the training period.

While quantitative measures of attention bias change (like d') are mentioned as being associated with BDI changes, specific pre- vs. post-training d' scores are not the primary focus of the reported results, which lean more towards the association between performance changes and mood changes.

Limitations and Future Work

The authors are candid about the limitations of their pilot study:

  • Lack of a Control Group: The most significant limitation is the absence of a control group. This means the observed improvements could be due to factors other than the neurocognitive therapeutic intervention, such as practice effects, placebo effects, or the natural course of the illness.
  • Small Sample Size: The study involved only seven participants, limiting the generalizability of the findings.

Future work should focus on:

  • Optimizing the Approach: Further refinement of the feedback regimes and task parameters.
  • Validating Effectiveness: Conducting larger, randomized controlled trials with appropriate active control groups to rigorously test the efficacy of neurocognitive therapeutics.
  • Expanding Scope: Applying this framework to other mood disorders and potentially other mental health conditions where attentional biases play a role.
  • Technological Translation: Exploring the transition from fMRI to more cost-effective, field-deployable neuroimaging technologies (like EEG) for broader accessibility.

Summary updated

Abstract

There is growing interest in the use of neuroimaging for the direct treatment of mental illness. Here, we present a new framework for such treatment, neurocognitive therapeutics. What distinguishes neurocognitive therapeutics from prior approaches is the use of precise brain-decoding techniques within a real-time feedback system, in order to adapt treatment online and tailor feedback to individuals' needs. We report an initial feasibility study that uses this framework to alter negative attention bias in a small number of patients experiencing significant mood symptoms. The results are consistent with the promise of neurocognitive therapeutics to improve mood symptoms and alter brain networks mediating attentional control. Future work should focus on optimizing the approach, validating its effectiveness, and expanding the scope of targeted disorders.
BibTeX
@inproceedings{schnyer2015neurocognitive,
  author = {David M. Schnyer and Christopher G. Beevers and Megan T. deBettencourt and Stephanie M. Sherman and Jonathan D. Cohen and Kenneth A. Norman and Nicholas B. Turk‐Browne},
  title = {{Neurocognitive therapeutics: from concept to application in the treatment of negative attention bias}},
  booktitle = {Biology},
  year = {2015},
  doi = {10.1186/s13587-015-0016-y},
}
Figure